The Impact of Wettability on Waterflooding: Pore-Scale Simulation

Abstract

Summary This paper describes the development and implementation of a pore-scale simulator into which pore-wettability effects have been incorporated. Relative permeability and capillary pressure curves from this steady-state model have been analyzed to allow better interpretation of experimental observations from a microscopic stand-point. The simulated capillary pressure data demonstrate that some standard wettability tests (such as Amott-Harvey and free imbibition) may give misleading results when the sample is fractionally wet in nature. Waterflood displacement efficiencies for a range of wettability conditions have been calculated, and recovery is shown to be maximum when the oil-wet pore fraction approaches 0.5. Furthermore, a novel experimental test is proposed that can be used to distinguish between fractionally wet and mixed-wet porous media. To date, no such satisfactory test exists. Introduction The wettability characteristics of a porous medium play a major role in a diverse range of petrophysical measurements that include capillary pressure data, relative permeability curves, and waterflood recovery efficiency. Throughout the current literature, however, much conflicting evidence exists as to the effects of wettability on a variety of displacement processes. Instead of examining the many factors that affect the ultimate wettability state of a system, however, this paper concentrates on the expected waterflooding consequences once a certain wettability condition has been specified at the pore scale. In any given rock sample, wettability may be considered as being either uniform or nonuniform. In the former case, the wettability of the entire pore space is the same (100% oil-wet, 100% water-wet, or 100% "intermediate"-wet) and the wetting- and nonwetting-phase contact angles will remain essentially constant throughout the system. In the more realistic nonuniform case, however, the rock may show heterogeneous wettability, with variations in wetting preference from pore to pore (e.g., 70% oil-wet pores and 30% water-wet pores). Indeed, wettability discontinuities may exist even at the subpore level, but such cases will not be considered here. Nonuniform wettability can be further subdivided into two subclasses: fractional wettability and mixed wettability. Fractional wettability is generally related to the rock matrix itself and is the result of differences in surface chemistry of the constituent minerals. Because of these variations, crude-oil components may adsorb onto some pore walls while ignoring others. In effect, this means that fractionally wet rock may contain oil-wet pores of all sizes (Fig. 1). Salathiel1 first introduced the term "mixed" wettability to describe systems where the oil-wet pores correspond to the largest in the sample, with the small pores remaining water-wet (Fig. 2). Such situations may arise when oil migrates to water-wet reservoirs and preferentially fills the larger interstices. The wettability characteristics of these pores may then be altered by adsorption of polar compounds and/or by deposition of organic matter from the original crude, thereby rendering them oil-wet. The network modeling approach has been widely used in the past, but usually in an attempt to quantify flow parameters2,3 or to fit experimental data.4 Unfortunately, quantitative prediction has proved unsuccessful, while the fitting of experimental data is less informative. We believe that network modeling should be used to examine the sensitivities of a given process to a variety of phenomena; this is the approach taken here. Although much work has been done with strongly water-wet networks, very few studies have been reported that attempt to simulate flow in networks of heterogeneous wettability. Mohanty and Salter5 have used a pore-and-throat model to investigate the effect of wettability on a variety of primary and secondary displacements but considered only one mixed-wet regime (25.5% PV oil-wet). In another study,6 a Bethe lattice was used to approximate pore connectivity and displacements were described usefully in terms of pore-filling sequences (schematic representations that use the pore-size-distribution curve of the system). Two nonuniform networks were studied, both of which appear to be fractionally wet. The simulator described in this work is used to investigate systematically the vital role played by wettability at the pore scale in both fractionally- and mixed-wet media and to explain how wettability influences the resulting capillary pressure, relative permeability, and waterflood recovery efficiency. In addition to explaining certain wettability sensitivities, an important goal of this work is to propose a new wettability test that eventually could be used in the laboratory. Model Setup and Implementation Network Details. The porous medium is modeled as an interconnected network of pore elements. The results presented in this paper come from simulations performed on a 3D regular cubic network comprising 20×20×20 nodes (junctions) that contains 24,400 intersecting pore elements. To simulate larger systems, periodic boundary conditions are imposed across planes perpendicular to the direction of the applied pressure gradient. Each pore in the network is assigned a radius, f(r), from a modified Rayleigh distribution. Network Details. The porous medium is modeled as an interconnected network of pore elements. The results presented in this paper come from simulations performed on a 3D regular cubic network comprising 20×20×20 nodes (junctions) that contains 24,400 intersecting pore elements. To simulate larger systems, periodic boundary conditions are imposed across planes perpendicular to the direction of the applied pressure gradient. Each pore in the network is assigned a radius, f(r), from a modified Rayleigh distribution.

Wettability and Imbibition: Microscopic Distribution of Wetting and Its Consequences at the Core and Field Scales

“…Insight into displacement mechanisms is provided by increasingly versatile deterministic network models (84)(85)(86)(87)(88). There is growing interest in the development of rules for pore level displacement that apply to multiphase flow.…”

Review of studies on pore-network modeling of wettability effects on waterflood oil recovery

“…Pore-network modeling has advanced to a veritable theoretical compliment to special core analysis and has proved to be an efficient tool to study the effect of wettability at the macroscopic scale, especially on oil recovery (McDougall and Sorbie, 1995;Blunt, 1998;DiCarlo et al, 2000;Zhao et al, 2010;Ryazanov et al, 2014). It is being used in the industry to assign multiphase flow properties in reservoir models and to predict how these properties vary with rock type and wettability.…”

“…It has been shown that temperature and oil viscosity also affect the distribution of fluids in the pore spaces, and hence wettability distribution (Alhammadi et al, 2017). McDougall and Sorbie (1995) investigated trends in relative permeability and oil recovery efficiency as functions of wettability with a regular cubic network. Recovery was shown to be maximum in a network where half the pore space was oil-wet.…”

3D Stochastic Modelling of Heterogeneous Porous Media – Applications to Reservoir Rocks

“…Another approach is to ignore the detailed structure of the porous medium and simply to represent the microstructure by an interconnected network based on the assumption that larger pores (pore bodies) are connected by smaller pores (throats) (Fatt, 1956;Kantzas et al, 1988;Øren and Pinczewski, 1992;McDougall and Sorbie, 1995;Blunt et al, 1995;Keller et al, 1997;Øren et al, 1994;Pereira et al, 1996;Mani and Mohanty, 1998;van Dijke and Sorbie 2002). Within such network models, it is possible to define a distribution of pores and throats so that the model reproduces various aspects of two-and three-phase flow behaviour.…”